Question bank

4,310 interview questions, answered.

Reference answers, what the interviewer is really testing, how it is graded, and the follow-ups that come next.

Easy 0Medium 24Hard 4
Filters · onDifficulty, topic, company

Difficulty

Clear all

71 questions

  1. 1Longest substring without repeating charactersAI ML· Amazon, Meesho
  2. 2Coding pairing: cache / tokenizer / reservation serviceMediumAI ML
  3. 3In-memory database with basic SQLMediumAI ML
  4. 4LRU cache — classic but dangerousMediumAI ML
  5. 5Most important OpenAI ship of the past yearMediumAI ML
  6. 6Reservation service under contentionMediumAI ML
  7. 7Streaming tokenization bottleneckMediumAI ML
  8. 8The mission-contradiction testHardAI ML
  9. 9Thread-safe cache with configurable evictionMediumAI ML
  10. 10Walk me through your favorite paperMediumAI ML
  11. 11Webhook delivery system (the paid work trial)MediumAI ML
  12. 12Word search on grid (Trie + DFS)MediumAI ML
  13. 13AI-enabled coding roundMediumAI ML
  14. 14AI safety trade-offs in your workMediumAI ML
  15. 15Inference serving under strict latency budgetsHardAI ML
  16. 16Longest path in an experiment-dependency graphMediumAI ML
  17. 17Model behavior that poses risk (signature question)HardAI ML
  18. 18The 15-minute phone screenMediumAI ML
  19. 19The most pressing unsolved alignment problemHardAI ML
  20. 20Transformer architecture & multi-head attentionMediumAI ML
  21. 213-hour technical assessment: BERT datasetAI ML
  22. 22Acceptable deployment trade-offsAI ML
  23. 23AI-assisted coding policy (verified current state)AI ML
  24. 24AI system with limited computeAI ML